trend-micro-ctf-2015miscchaitan94romanparsing
Again, problem is simple: We're given simple arithmetic questions, we need to calculate their values.
In the start, it's quite simple as the questions contained just numbers and operators. Python's eval() would've been more than enough, but as level increase, we get numbers separated by commas (like 62,234,612). Even more, we start to see roman numerals (CMXIV), and finally even numbers in plain english! (six hundred twenty two).
To automate this, we wrote a python script which after reading each question, does the following:
def solve(ques):
# We only need lhs
ques = ques.split("=")[0]
# We don't need commas
ques = ques.repace(',', '')
parsed = ''
# Let's separate parts between operators
groups = re.split('([\-\+\*\/\(\)])', ques)
# Convert each part to a number
for g in groups:
g = g.strip()
# If part a roman numeral, convert it to decimal
if re.match('([A-Z]+)', g):
parsed += "%s" % roman.fromRoman(g)
# If part is in plain english, convert it to decimal
elif re.match('([a-z]+)', g):
parsed += "%s" % text2int(g)
else: parsed += g
# Now eval can do the rest
return "%s\r\n" % (eval(parsed))(Complete code is in bot.py)
After about 100 levels, we get the flag:
TMCTF{U D1D 17!}